Metadata: Affordances and Constraints

by Curtis Hisayasu

Click on the map to read more about the synthesis layer.

To me, the most productive aspect of the “Standards in the Making” (SITM) approach which we have, collectively, been attempting to elaborate, is its particular alignment of critical and creative practice.  In this entry, I would like to try and understand the relationship between SITM’s status as a critical lens on metadata and its emphasis on “making standards” as a part of the new work of digital composition.  For my part, this involves returning to and reconsidering what Matt calls “the multiple affordances and constraints in making media.”  On this topic, our collective investigation of the Flickr Commons Project reveals a number of thematic intersections that I would like to think through here.  In particular, I would like to consider how the actual practice of “standards in the making” takes up the various “constraints” of folksonomic metadata systems as “affordances” themselves—how composing metadata can “afford” an engagement and an intervention into “constraints,” variously understood as emerging from a variety of overlapping contexts.  Conversely, I would like to argue that the limits imposed upon the production of metadata are themselves generative constraints that facilitate (as well as restrict) the way that metadata itself operates as a compositional practice.  At the intersection of our individual analyses, a fuller and more complex understanding of the “situation” of metadata and, consequently, an opportunity for “new work” in digital composition, begins to emerge.

Constraints: The Situation of Metadata

One such intersection concerns the ultimate “predictability” of the user contributions to the Commons project.  Our collective investigation of this archive inevitably ran up against questions regarding the predictable nature of the “standards” that emerge within the archive as a consequence of user generated tags and comments, and a good deal of our analysis was expended attempting to suggest which types of “constraints” were responsible for this standardization.  Even metadata that emerges from folksonomic tagging systems, as it turns out, is not so distributed, decentralized, or thoroughly individualized that it cannot be understood as being subject to various forms of ideological and structural “constraint.”  While this seems to be a point of consensus across our responses to the LoC archive, it is worthwhile to take stock of the variety of contexts that can be said to “constrain” the production of metadata, and consequently the production of the archive itself.  In terms of “standards in the making” as a practice of composing in new media, these contexts themselves constitute an important part of what Jamie calls “the rhetorical situation of metadata” which is essential to understanding its status as a discursive space where “arguments are mobilized through the metadata itself.”

I would like to imagine these different contexts as overlapping sites of power/knowledge, which quietly and surreptitiously standardize, if not the specific content of the metadata itself, then the limits and possibilities for categorization and labeling.  Jentery’s assessment of the Springer report, for instance, suggests that around 45% of user tagging is “redundant” and derivative of Library of Congress’ own record categories—suggesting, at the very least, that these users are somehow invested in reproducing official institutional practice and vocabulary.  Matt, in a different vein, worries that “the coding of these annotations in the LoC Flickr Commons have become too crystallized, too stabilized, too group-think” in the way that the very “possibility for annotations have been prefigured into bins for tags, notes, and comments.”  By standardizing the available “forms of documentation,” Flickr’s user interface is therefore able to centralize standards for “how those tags, notes, and comments should be appropriately composed.” Another site of constraint then is located at the level of the software infrastructure, where the centralized organization of user contributions has already structured the “uses” that those contributions can be put to.  But yet another site of constraint might be located in the intersection of social tagging practices and larger social knowledge projects.  My own analysis was dedicated to illustrating how the construction of so-called descriptive metadata around racial identity is actually caught up in (and, to a certain extent, determined by) the contradictions and questions of race that circulate generally in our contemporary moment.  Through this analysis, I attempted to suggest that another site of constraint for the production of social metadata is in the terminological complexities and representational logics that make up our contemporary social formation.

There are, of course, methodological and conceptual conflicts between these readings that we could attend to, but for my purposes here I would simply like to call attention to how these contributions outline the variety of means in which the context of metadata can be understood.  Each of these different (but only provisionally discrete) “contexts”—the institutional, the infrastructural, the social – may be considered a part of the situation which discretely standardizes the production of social metadata.  As Jentery notes, the ultimate goal of SITM is not simply in “explaining away” the predictable “recapitulations of already existing cultural norms and logics.”  Here, he invokes the concept of “implicit metadata” which suggests (or, perhaps, demands) that we inquire into the all of these different sites of constraint in order to discover what types of embedded “standards are fostering the co-production of metadata in the first place.”  Seen through one lens, this call (implicit, in some ways, in all of our responses to the LoC project) describes SITM’s function as a critical practice – an analytic that seeks to diagnose and comprehend the relations of power that circulate in the unspoken contexts of metadata’s composition.

Affordances: The Opportunities of SITM

How to transition to “affordances” and, in doing so, consider more directly the work of SITM in the creation of digital media (rather than in the critique of folksonomies)?  If we want to understand the project of SITM as a component of “the new work of composing” then we must, as I argued earlier, turn to considering these contexts as a part of the “rhetorical situation” of metadata.  The first component of this work, as Jamie notes, is to understand the work of metadata as inherently “argumentative, and not merely descriptive.”  Amongst other things, I take this as a call to understand each of these forms of constraint as potentially generative – as a part of the infrastructural apparatus through which meanings are produced and communicated (as well as limited and regulated).  Indeed, in the rhetorical situation defined by the analyses synthesized above, each instance of metadata assumes the value and meaning it does precisely because of the way that it has been “constrained” in each context.  Here I find myself moving contrary to Matt’s suggestion that SITM implies a way of “producing standards without confronting the horizon of the Standard.”  What I am attempting to argue is that it is precisely by confronting the various limits and assumptions of “the Standard” that “making standards” can be a creative endeavor that simultaneously assumes a critical weight.  In this way, I will return to my argument about the role of metadata in the production of knowledge itself.

One of the advantages of envisioning the overlapping contexts of metadata as a “rhetorical situation” is that it forces us to understand the practice of social tagging as an inherently conversational one.  Rather than envisioning metadata as either a non-hierarchical “piling on” of equally-valid descriptions and search terms in the “flat” space of the Commons, the critical practice of SITM outlined above depends upon seeing user tags in dialogue with one another (both Jentery and I note in our individual analyses that the terms that qualify race in the Commons archive overlap and should be read as competing arguments about the nature of racial identity itself).  Additionally, metadata contributed even under these auspices are often not simply arguments about the status of the object in question, but often quite visibly constitute attempts to navigate the very constraints that qualify and define appropriate discourse within each situation.  The brief conversation about what does or doesn’t constitute “useful metadata” in the notes section of “Children asleep on a bed…” is a relatively low-stakes example of this.  Understanding how, as Matt’s analysis suggests, these notes “exist to motivate a ‘common understanding’” is vitally important, but so is the potential for “problematization,” where unexpected tags can intervene, differently navigate, or disrupt consensus on and constraints around what constitutes “useful” metadata—indeed what the “use” of metadata is and should be.

Of course, the discursive affordances of social metadata should not be simply understood as enabling another conversational scenario like any other; the preceding analyses insist that metadata’s “situation” clearly has its own infrastructural and compositional constraints.  Also, I might return to my earlier thesis (that tags are active in producing and defining objects as specific forms of knowledge) in order to insist that the real rhetorical “work” of metadata is not how it “describes” individual objects, but how it organizes and consolidates the archive in idiosyncratic and argumentative ways.  I would argue that it is here that the real “conversation” occurs – in the way that the arguments that metadata stages are not simply about what the object “is” but what it belongs to, in what sequence, according to what logics, and in relation to what knowledge projects.  SITM, as a practice of “making standards” to see what emerges, is a way of imposing dialogue and problematics within the logic of the archive.  We can only speculate, for instance, what kinds of arguments a new metadata category might make in relation to both its competing tagged categories and to the standards imposed by its context, variously understood.  How might the term “Black Labor” for instance cut across other archival logics, variously intersecting and digressing from the terms of identity that define the categories of “Negro,” “Black,” and “African-American” in the earlier “Negro boy…” and “Rosie” examples?  How might it argue for a different vision of black history and identity against the Standards defined by institutional or social formations – one that, for instance, draws a much different continuity through pre-emancipation, Jim-Crow, and contemporary contexts?  If composing and mobilizing such a standard is to have any meaning—if it is to achieve the status of an argument – it will construct that meaning, in part, in relation to the other standards that also compete to define the archive and its variously captured objects.  Tagging “composes” the archive and, in doing so, dialogues with other metadata arguments about how that same archive can and should be understood.  Working very much within the constraints of each metadata situation (in some cases, poaching their logics and proscribed “uses”), SITM locates and writes through the opportunities afforded by the tangibly fragile nature of “the Standard’ itself.